AI is saving time and money in research – but at what cost? (www.nature.com)

A recent survey by Wiley reveals a significant uptick in the use of AI tools by researchers, with 62% of over 2,400 respondents utilizing AI for tasks like writing, editing, and error detection, up from 45% in 2024. Particularly popular among early-career scientists and those in physical sciences, AI has been praised for enhancing efficiency and output, with 85% of users reporting increased productivity and quality in their work. For instance, astrophysicist Matthew Bailes notes that AI assists in processing vast datasets in astronomy, streamlining the identification of neutron star signatures and facilitating innovative educational simulations of cosmic phenomena. While the productivity gains from AI are evident, there are growing concerns regarding its implications for scientific integrity and diversity. The same survey highlighted that 87% of respondents are worried about the potential for AI-generated errors, often referred to as hallucinations, along with issues related to data security and the ethics of AI applications. As AI continues to accelerate research outputs—evidencing a greater number of publications and citations—it raises critical questions about the balance between technological advancement and maintaining diverse, error-free scientific discourse.